: Accelerometry has become the method of choice for measuring physical activity in free living children and adolescents. However, despite the widespread use of accelerometers in pediatric research, interpreting the output from these devices is problematic. Parameters related to normal growth and development in children (i.e., increasing stature) influence the magnitude of activity counts recorded by an accelerometer, making it difficult to directly compare count indices in children of different ages. In addition, developmental differences with respect to resting metabolic rate and economy of movement mandate that equations or algorithms to convert raw accelerometer counts to units of energy expenditure and/or physical activity intensity, take into consideration chronological age or other practical indicators of growth and development. Although existing calibration equations/algorithms consider age-related differences in energy metabolism, they were developed from cross-sectional evaluations of accelerometer counts and energy expenditure. Thus, the longitudinal validity of these equations/algorithms, and hence their true utility in longitudinal studies has not been investigated. A second important methodological issue is the ability of accelerometers to detect change in free-living physical activity behavior. Presently, we do not know if the existing equations or cut- points are sufficiently sensitive to detect the small changes in free-living activity behavior that are typically induced by intervention programs. The specific aims of this study are to: 1) evaluate and compare the longitudinal validity of previously published energy expenditure calibration equations/algorithms for the Actigraph, Actical, and RT3 accelerometers in an age-diverse cohort of children and adolescents (6-15 yrs);and 2) determine if accelerometers and their respective count cut-points are sufficiently sensitive to detect the changes in moderate-to-vigorous physical activity (MVPA) that would typically be induced by a physical activity intervention program. We will examine how normal growth and development influence accelerometer predictive validity and senstivity to change by conducting evaluations at 12-, 24- and 36 months follow-up. The results will enable researchers and practitioners to more effectively reduce and interpret accelerometer data collected in longitudinal intervention studies of youth as well as population-based studies involving age diverse samples of children and adolescents.